CN116434918A - Medical image processing method and computer readable storage medium - Google Patents
Medical image processing method and computer readable storage medium Download PDFInfo
- Publication number
- CN116434918A CN116434918A CN202111652660.4A CN202111652660A CN116434918A CN 116434918 A CN116434918 A CN 116434918A CN 202111652660 A CN202111652660 A CN 202111652660A CN 116434918 A CN116434918 A CN 116434918A
- Authority
- CN
- China
- Prior art keywords
- medical image
- medical
- processing method
- artifact
- image processing
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000003672 processing method Methods 0.000 title claims abstract description 38
- 238000013441 quality evaluation Methods 0.000 claims abstract description 31
- 238000012216 screening Methods 0.000 claims abstract description 30
- 238000000034 method Methods 0.000 claims description 17
- 238000011156 evaluation Methods 0.000 claims description 13
- 238000010801 machine learning Methods 0.000 claims description 6
- 238000007781 pre-processing Methods 0.000 claims description 6
- 238000001914 filtration Methods 0.000 claims description 3
- 230000005856 abnormality Effects 0.000 claims 2
- 238000001514 detection method Methods 0.000 abstract description 8
- 238000004364 calculation method Methods 0.000 abstract description 5
- 210000003739 neck Anatomy 0.000 description 11
- 238000013527 convolutional neural network Methods 0.000 description 8
- 239000002184 metal Substances 0.000 description 7
- 238000001303 quality assessment method Methods 0.000 description 6
- 238000013528 artificial neural network Methods 0.000 description 4
- 210000000746 body region Anatomy 0.000 description 4
- 210000004556 brain Anatomy 0.000 description 4
- 238000003384 imaging method Methods 0.000 description 4
- 210000000056 organ Anatomy 0.000 description 4
- 230000011218 segmentation Effects 0.000 description 4
- 238000012549 training Methods 0.000 description 4
- 238000004891 communication Methods 0.000 description 3
- 230000000694 effects Effects 0.000 description 3
- 238000012545 processing Methods 0.000 description 3
- 240000004282 Grewia occidentalis Species 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 2
- 210000005013 brain tissue Anatomy 0.000 description 2
- 238000003759 clinical diagnosis Methods 0.000 description 2
- 238000003745 diagnosis Methods 0.000 description 2
- 238000002059 diagnostic imaging Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 238000001208 nuclear magnetic resonance pulse sequence Methods 0.000 description 2
- 230000000241 respiratory effect Effects 0.000 description 2
- 238000005070 sampling Methods 0.000 description 2
- 238000000264 spin echo pulse sequence Methods 0.000 description 2
- 230000004075 alteration Effects 0.000 description 1
- 210000000481 breast Anatomy 0.000 description 1
- 230000015556 catabolic process Effects 0.000 description 1
- 238000007405 data analysis Methods 0.000 description 1
- 238000013523 data management Methods 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000006731 degradation reaction Methods 0.000 description 1
- 230000006866 deterioration Effects 0.000 description 1
- 238000003748 differential diagnosis Methods 0.000 description 1
- 201000010099 disease Diseases 0.000 description 1
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 1
- 229940079593 drug Drugs 0.000 description 1
- 239000003814 drug Substances 0.000 description 1
- 230000001490 effect on brain Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 210000004884 grey matter Anatomy 0.000 description 1
- 239000007943 implant Substances 0.000 description 1
- 210000003734 kidney Anatomy 0.000 description 1
- 230000003902 lesion Effects 0.000 description 1
- 210000003141 lower extremity Anatomy 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000003062 neural network model Methods 0.000 description 1
- 230000007170 pathology Effects 0.000 description 1
- 230000008855 peristalsis Effects 0.000 description 1
- 230000010349 pulsation Effects 0.000 description 1
- 239000000700 radioactive tracer Substances 0.000 description 1
- 230000000306 recurrent effect Effects 0.000 description 1
- 230000029058 respiratory gaseous exchange Effects 0.000 description 1
- 230000001629 suppression Effects 0.000 description 1
- 230000002792 vascular Effects 0.000 description 1
- 210000004885 white matter Anatomy 0.000 description 1
- 238000004804 winding Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H30/00—ICT specially adapted for the handling or processing of medical images
- G16H30/20—ICT specially adapted for the handling or processing of medical images for handling medical images, e.g. DICOM, HL7 or PACS
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G06T5/70—
-
- G06T5/73—
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0012—Biomedical image inspection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/11—Region-based segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/136—Segmentation; Edge detection involving thresholding
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10072—Tomographic images
- G06T2207/10081—Computed x-ray tomography [CT]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10072—Tomographic images
- G06T2207/10088—Magnetic resonance imaging [MRI]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20024—Filtering details
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20081—Training; Learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20084—Artificial neural networks [ANN]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20112—Image segmentation details
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/30—Computing systems specially adapted for manufacturing
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Radiology & Medical Imaging (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Biomedical Technology (AREA)
- Data Mining & Analysis (AREA)
- Primary Health Care (AREA)
- Quality & Reliability (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Public Health (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Epidemiology (AREA)
- Evolutionary Computation (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Apparatus For Radiation Diagnosis (AREA)
Abstract
The invention provides a medical image processing method and a computer readable storage medium, comprising the following steps: acquiring a plurality of medical images to be detected; performing preliminary screening on a plurality of medical images to determine candidate medical images; and carrying out image quality evaluation on the candidate medical images to obtain quality evaluation results. According to the medical image processing method provided by the invention, before the quality evaluation of the medical image, the candidate medical image is determined through preliminary screening, so that the image quality can be rapidly screened, the calculation force is saved, and the detection efficiency of the image quality is improved.
Description
Technical Field
The present invention relates to the field of medical image processing technology, and in particular, to a medical image processing method and a computer readable storage medium.
Background
In medical imaging systems, image quality depends on many factors, such as spatial resolution, tissue contrast, signal to noise ratio, contrast to noise ratio, image defects. In order to exhibit the best image quality, both hardware and scan parameters are optimized for different organs or pathologies. In scanning, however, due to physiological factors of the subject (respiration, heartbeat, body structure), the movement of the subject at the time of scanning may cause deterioration of image quality, thereby failing to meet clinical diagnosis requirements.
To ensure scan quality, in scanning, the user needs to manually retrieve and view the scanned images, evaluate the integrity of the information contained in the images to ensure that they are of acceptable quality prior to processing the data analysis, and determine whether a re-scan is required to ensure the need. Particularly for whole-body imaging, the whole-body image acquisition can only be completed in a mode of respectively acquiring a plurality of beds due to the limitation of hardware, and the conventional magnetic resonance scanning on each bed needs to comprise images with different weights, such as T1, T2 and DWI, and different azimuth acquisitions of the images with the same weights, such as T2 transverse position, T2 coronal position, different parameter acquisitions and the like. In addition, the differential diagnosis can be performed by scanning more specific magnetic resonance sequences at certain positions aiming at different diseases. There are thus at least twenty sequences that ultimately lead to MR routine examinations, when the image is viewed for image quality. This procedure is time consuming and laborious, certainly adding to the burden on the physician.
Artifacts (Artifacts) are images of various forms that appear on an image without the existence of an object to be scanned, and are important factors that cause the quality degradation of medical images, and even greatly affect the analysis and diagnosis of lesions by doctors. Therefore, as a medical diagnosis basis, medical image artifact identification is of great importance. Traditional methods of learning (machine/deep learning) can be used to predict artifacts, the higher the accuracy, the larger the model and the greater the computational effort required by the computer. In practical application, the image quality problem is less than 5%, if the image quality analysis operation amount of each scanned medical image is huge, the operation of other algorithms can be influenced. Therefore, it is necessary to provide a rapid image quality screening method to save the calculation effort and improve the detection efficiency of the image quality.
Disclosure of Invention
The invention aims at a medical image processing method and a computer readable storage medium to realize quick screening of image quality, save calculation force and improve the detection efficiency of the image quality.
To achieve the above object, the present invention provides a medical image processing method, comprising:
acquiring a plurality of medical images to be detected;
performing preliminary screening on a plurality of medical images to determine candidate medical images;
and carrying out image quality evaluation on the candidate medical images to obtain quality evaluation results.
Optionally, performing a preliminary screening of the plurality of medical images to determine candidate medical images includes:
segmenting a tissue region in at least one of the medical images;
setting a first threshold and a second threshold according to the segmented medical image respectively, and calculating an evaluation parameter, wherein the evaluation parameter is the ratio of the number of pixels with pixel values between the first threshold and the second threshold in the medical image to the number of pixels in the tissue region;
and performing preliminary screening on the medical images according to the relation between the evaluation parameters and a set threshold value to obtain candidate medical images.
Optionally, a maximum pixel value in the background of the medical image is extracted as a first threshold value, and a minimum pixel value in the tissue region is extracted as a second threshold value.
Optionally, the identification of the body part in the medical image is also included before segmenting the tissue region in the medical image.
Optionally, the identifying of the body part in the medical image may further comprise pre-processing the medical image, the pre-processing including down-sampling, filtering or normalizing.
Optionally, when the evaluation parameter is greater than the set threshold, the medical image is determined to be the candidate medical image.
Optionally, a trained neural network model is employed to evaluate the quality of the candidate medical image, the quality characterized by at least one image attribute parameter of image integrity, image contrast, image signal-to-noise ratio, image resolution.
Optionally, the quality assessment results include normal, general, moderate anomaly, or severe anomaly.
Optionally, the medical image is obtained by scanning a scanning object by a scanner, and the medical image processing method further includes: and determining the state parameters of the scanner according to the quality evaluation result.
The present invention also provides a computer readable storage medium storing at least one instruction executable by a processor, the at least one instruction, when executed by the processor, implementing the medical image processing method according to any one of the above.
In summary, the present invention provides a medical image processing method and a computer readable storage medium, including: acquiring a plurality of medical images to be detected; performing preliminary screening on a plurality of medical images to determine candidate medical images; and carrying out image quality evaluation on the candidate medical images to obtain quality evaluation results. According to the medical image processing method provided by the invention, before the quality evaluation of the medical image, the candidate medical image is determined through preliminary screening, so that the image quality can be rapidly screened, the calculation force is saved, and the detection efficiency of the image quality is improved.
Drawings
FIG. 1 is a flowchart of a medical image processing method according to an embodiment of the present invention;
FIG. 2 is a flowchart illustrating a process for acquiring a candidate medical image in a medical image processing method according to an embodiment of the present invention;
FIG. 3 is a schematic diagram illustrating a segmentation of a non-tissue region and a tissue region in a medical image according to an embodiment of the present invention;
FIG. 4 is a block diagram of a convolutional neural network of two layers of CNNs used for artifact quality assessment in a medical image processing method according to an embodiment of the present invention;
FIG. 5 is a flowchart of establishing an artifact screening network model in a medical image processing method according to an embodiment of the present invention;
FIG. 6 is a flowchart of a medical image processing method according to another embodiment of the present invention;
fig. 7 is a flowchart of a process for obtaining a quality evaluation result in a medical image processing method according to an embodiment of the present invention.
Detailed Description
The medical image processing method and the computer-readable storage medium of the present invention are described in further detail below with reference to the accompanying drawings and detailed embodiments. The advantages and features of the present invention will become more apparent from the following description and drawings, however, it should be understood that the inventive concept may be embodied in many different forms and is not limited to the specific embodiments set forth herein. The drawings are in a very simplified form and are to non-precise scale, merely for convenience and clarity in aiding in the description of embodiments of the invention.
The terms "first," "second," and the like in the description are used for distinguishing between similar elements and not necessarily for describing a particular sequential or chronological order. It is to be understood that the terms so used are interchangeable under appropriate circumstances such that the embodiments of the invention described herein are, for example, capable of operation in other sequences than described or illustrated herein. Similarly, if a method described herein comprises a series of steps, and the order of the steps presented herein is not necessarily the only order in which the steps may be performed, and some of the described steps may be omitted and/or some other steps not described herein may be added to the method. If a component in one drawing is identical to a component in another drawing, the component will be easily recognized in all drawings, but in order to make the description of the drawings clearer, the specification does not refer to all the identical components in each drawing.
The present embodiment provides a medical image processing method, fig. 1 is a flowchart of the medical image processing method provided in the present embodiment, and as shown in fig. 1, the medical image processing method provided in the present embodiment includes:
step S11: acquiring a plurality of medical images to be detected;
step S12: performing preliminary screening on the plurality of medical images to determine candidate medical images;
step S13: and performing image quality evaluation on the candidate medical images to obtain quality evaluation results.
Alternatively, the plurality of medical images may be reconstructed in real time by a scanner of the medical device or obtained by a medical image data management system, such as a Picture Archiving and Communication System (PACS). Alternatively, the medical image processing method may be performed by a post-processing software. Alternatively, a plurality of medical images may be stored in a file in accordance with the medical communication standard DICOM. The DICOM standard (Digital Imaging and Communication in Medicine ) standardizes the structure of the format and descriptive parameters for radiological images and commands for exchanging these images, as well as other data objects such as image sequences, examination sequences and examination reports.
Alternatively, the determination process of the candidate medical images may be equivalent to a process of preliminary recognition of a plurality of medical images. Alternatively, the preliminary screening of the plurality of medical images may be performed by calculating the resolution of each medical image, by calculating the signal-to-noise ratio of each medical image, by calculating the contrast of each medical image, or by calculating whether each medical image has a region of interest (ROI) to determine whether the image is complete.
Optionally, performing a preliminary screening of the plurality of medical images to determine candidate medical images includes: segmenting a tissue region in the at least one medical image; setting a first threshold and a second threshold according to the segmented medical image respectively, and calculating an evaluation parameter, wherein the evaluation parameter is the ratio of the number of pixels with pixel values between the first threshold and the second threshold in the medical image to the number of pixels in a tissue region; and performing preliminary screening on the plurality of medical images according to the relation between the evaluation parameters and a set threshold value to obtain candidate medical images. The first threshold is different from the second threshold, for example: the first threshold may be determined from pixel values of pixels comprised by a background (e.g. a non-tissue region) of the medical image and the second threshold may be determined from pixel values of pixels comprised by a tissue region in the medical image. The evaluation parameter in the embodiments of the present application may be a medical image quality parameter. The evaluation parameter may be an artifact parameter, a resolution parameter, a contrast parameter, or an integrity parameter, etc. Of course, the first threshold value and the second threshold value may also be set by empirical values, which is not limited in comparison in the embodiment of the present application.
Alternatively, the quality evaluation result of the candidate medical image may be, for example, the degree/level of the artifact appearing in the medical image, the contrast level corresponding to the medical image, whether the medical image is applicable/usable for the current clinical diagnosis, or the like. Optionally, the artifacts in the medical image may include: motion artifacts generated by autonomous motion and convolution artifacts generated by phase winding; gradient spark artifacts; the detection object carries metal artifacts generated by metal; scattering shading-like artifacts; artifacts from contrast/tracer; truncation artifacts; respiratory motion artifacts, and the like.
In one embodiment, the candidate medical image is obtained by initially screening the medical image for the presence of motion artifacts that cause blurring of the image background. As shown in fig. 2, the acquisition of the candidate medical image includes:
step S21: acquiring a medical image to be detected;
step S22: segmenting a tissue region and a non-tissue region in at least one of the medical images;
step S23: extracting any pixel value of a set area of the background of the medical image as a first threshold value M1, and any pixel value of the tissue area as a second threshold value M2, and calculating an artifact parameter P, wherein the artifact parameter P is the ratio of the number M of pixels, of which the pixel value is between the first threshold value M1 and the second threshold value M2, in the medical image to the number N of pixels in the tissue area; the method comprises the steps of,
step S24: and judging that the medical image is free of artifacts according to the relation between the artifact parameter P and a set threshold value.
Specifically, first, a medical image to be detected is acquired. The medical image may be a multi-source medical image, i.e. from different types of medical images, such as CT, MRI, etc.; or medical images from different hospital equipment. For example, medical images are MRI, wherein the scan field strength comprises 1.5T (tesla), 3.0T, etc., the scan site comprises a head transverse, a head coronal, a head sagittal, a neck transverse, a neck coronal, a neck sagittal, etc., the scan direction comprises a coronal, a sagittal, a transverse, and the scan sequence comprises a gradient echo pulse sequence (GRE) or a fast spin echo pulse sequence (FSE). The graph screening method provided by the embodiment can support artifact identification of different parts, such as head, spine, lower limbs and the like; different artifact types are supported for identification, such as motion artifacts, metal artifacts, etc.
The medical image is then preprocessed. The preprocessing includes downsampling, filtering or normalizing. For example downsampling the medical image to a size of 320 x 320.
Next, a body part in the medical image is identified. For example, the division may be implemented according to the morphology of the human body region. Since human organs or tissues are in the form of, for example, the same general commonality: the kidney is bean-shaped and the breast is hilly. Correspondingly, the gray level distribution characteristics of organs or tissues in the human body region image are also in a certain preset range, and the non-tissue region can be divided from the human body region image based on the gray level distribution characteristics in the human body region image, so that the rapid screening of candidate images is facilitated.
Next, as shown in fig. 3, the tissue region and the non-tissue region in the medical image are segmented. For example, the tissue and non-tissue regions are segmented by thresholding, and the selection of the threshold during the thresholding of the tissue and non-tissue regions is determined based on the background noise/noise of the medical image. Whether the medical image is background noise or flat on the background noise determines whether the image processing-based artifact screening method is stable. In the pure background noise environment, even a simple threshold segmentation detection method can obtain a better artifact detection effect, however, in general, the obtained medical image signals have background noise. Therefore, the medical image processing method we establish must have a better robustness to noise. That is, the threshold value of the thresholding segmentation is determined based on the background noise of the medical image during segmentation of the tissue region and the non-tissue region. In addition, the background noise extraction modes are correspondingly different for the image data of medical images acquired by different acquisition modes.
Then, any pixel value of a set area of the background of the medical image is extracted as a first threshold value M1, any pixel value of the tissue area is extracted as a second threshold value M2, and an artifact parameter P is calculated, wherein the artifact parameter P is a ratio of the number M of pixels in the medical image, between the first threshold value M1 and the second threshold value M2, to the number N of pixels in the tissue area, i.e., p=m/N. Alternatively, the setting area of the background area may be an area containing four-corner information. The four-corner information may be, for example, display mode information, image attribute information, and the like included in the upper left corner, lower left corner, upper right corner, and the like of the medical image.
Next, referring to fig. 2, it is determined that the medical image is artifact free according to the relationship between the artifact parameter P and a set threshold. Specifically, the relationship between the artifact parameter P and a set threshold is determined, when the artifact parameter P is greater than a set threshold, the medical image is determined to have an artifact (output 1), and when the artifact parameter P is less than or equal to a set threshold, the medical image is determined to have no artifact (output 0).
Image quality evaluation is performed on the candidate medical images to obtain quality evaluation results, which can be achieved through a machine learning network model. The machine learning network model may be DNN (Deep Neural Networks, deep neural network), CNN (Convolutional Neural Networks, convolutional neural network), RNN (Recurrent Neural Network ), etc., and when the machine learning network model is CNN, it may be a V-Net model, a U-Net model, a generative countermeasure network Generative Adversarial Nets model, etc. Quality can be characterized by at least one image attribute parameter of image integrity, image contrast, image signal to noise ratio, image resolution.
In one embodiment, the quality is characterized by an image signal-to-noise ratio. When it is determined that an artifact is present in the medical image, an assessment of the quality of the artifact is also required. The quality of the artifact may be assessed using a neural network. Multiple classification artifact quality assessment is typically done using either Resnet 18 or Resnet 50, e.g., an artifact quality assessment may be done with a simple layer 2 CNN, as shown in FIG. 4. The result of the artifact quality assessment is expressed in terms of artifact levels, including normal, general, moderate or severe anomalies.
Still alternatively, in one embodiment, to facilitate determining a degree of impact of the artifact in the candidate medical image on the quality of the medical image to be processed, the quality assessment result output by the machine learning network model is artifact degree indication information. Optionally, the computer device may rank the artifact level indication information into a first level, a second level, a third level, and a fourth level, wherein the first level may represent that the candidate medical image is normal and not affected by the artifact; the second level may represent a slight effect of the artifact on the candidate medical image, which may continue to be used; the third level can represent the moderate influence of the artifact on the candidate medical image, and the artifact cannot be clinically used; level four indicates that the artifact has severe impact on the candidate medical image and cannot be used clinically.
The quality evaluation of the artifact may also be performed by an artifact screening network model, as shown in fig. 5, where the process of establishing the artifact screening network model includes:
step S51: collecting image data of medical images, constructing data sets of different scenes and different artifact degrees, and preprocessing to obtain a training data set;
step S52: constructing a part identification network, and obtaining the number of a scene corresponding to the medical image according to the image data;
step S53: constructing an artifact screening network model based on the deep convolutional neural network;
step S54: performing network training on the artifact screening network model according to the training data set and the serial numbers of the scenes, and calculating the artifact degree of the medical image through the artifact screening network model after the network training;
step S55: and properly adjusting the data set according to the artifact degree calculated by the artifact screening network model.
Wherein the scene in step S51 comprises a scan site comprising a head transverse, a head coronal, a head sagittal, a neck transverse, a neck coronal or a neck sagittal, and a scan sequence comprising a gradient echo pulse sequence (GRE) or a fast spin echo pulse sequence (FSE). The artifact level includes normal, general, moderate or severe anomalies.
And generating prompt information according to the quality evaluation result. In one embodiment, the computer device outputs the prompt if the degree of influence of the artifact in the medical image to be processed on the image quality of the candidate medical image is greater than or equal to a preset artifact influence degree threshold.
The prompting information can be a prompting identifier, and is used for prompting a user to confirm whether to accept the artifact of the candidate medical image and whether to need to re-scan the scanned part corresponding to the candidate medical image; or, the prompt information can be only a warning mark for indicating that the medical image to be processed has the artifact affecting the image quality; the hint information may be a specific sequence that corresponds to the medical image affected by the artifact and that is a time sequence throughout the medical imaging scan. The mode of outputting the prompt information by the computer equipment can also be to send out the prompt sound, can also be to send out the prompt red light, can also be to display the prompt text scanned again on the display screen, and the mode of outputting the prompt information by the computer equipment is not particularly limited in the embodiment of the application.
In one embodiment, the medical image is obtained by scanning and reconstructing a scanning object in real time by a scanner, and as shown in fig. 6, the medical image processing method includes:
step S61: acquiring a plurality of medical images to be detected;
step S62: performing preliminary screening on the plurality of medical images to determine candidate medical images;
step S63: performing image quality evaluation on the candidate medical images to obtain quality evaluation results;
step S64: and determining the state parameters of the scanner according to the quality evaluation result.
The state parameters of the scanner may be the operational state parameters of the device itself, such as gradient parameters, radio frequency parameters, main magnetic field parameters, etc.; the state parameter of the scanner may also be the state of the scanned object, e.g. whether metal is carried, whether autonomous motion is present. Optionally, the determining of the state parameter of the scanner may include: determining that the resolution of the candidate medical image does not reach a set value according to the quality evaluation result, setting the scanning time of the scanner to be too short, and needing to prolong the scanning time of the scanner or improve the sampling rate; or determining that the candidate medical image has spark artifact according to the quality evaluation result, and changing the gradient waveform; or, if the candidate medical image is determined to have metal artifact according to the quality evaluation result, the scanning object of the scanner carries metal, and an artifact suppression algorithm needs to be added.
In one embodiment, taking the example of the existence of motion artifact in the candidate medical image, as shown in fig. 7, the process of obtaining the quality evaluation result may include:
step S71, inputting the candidate medical image into the target artifact identification model to obtain target artifact attribute information output by the target artifact identification model.
After the target artifact identification model identifies the candidate medical image, target artifact attribute information may be output, where the target artifact attribute information is used to indicate attribute characteristics of artifacts in the candidate medical image. The target artifact attribute information may include at least one of size information of the artifact, position information of the artifact, number information of the artifact, and category information of the artifact. The types of artifacts may include zipper artifacts, spark artifacts, involuntary motion artifacts, respiratory artifacts, vascular pulsation artifacts, and the like, among others. The types of artifacts can also be classified into equipment artifacts and artificial artifacts according to sources, and the equipment artifacts comprise measurement error artifacts of an imaging system, X-ray beam hardening artifacts, high-voltage fluctuation artifacts of the imaging system, temperature drift artifacts of an electronic circuit, detector drift artifacts and the like; artifacts include, for example, artifacts that detect subject body position movement, artifacts that peristalsis of organs in the body, artifacts of metal implants in the body, and the like.
Step S72, inputting the candidate medical image and the target artifact attribute information into the target artifact degree identification model to obtain artifact degree indication information output by the target artifact degree identification model.
Wherein the artifact level indication information is used to indicate a degree of influence of artifacts in the candidate medical image on the image quality of the candidate medical image.
Specifically, after inputting the candidate medical image to the target artifact identification model and obtaining the target artifact attribute information output by the target artifact identification model, the computer device may input the candidate medical image and the target artifact attribute information to the target artifact extent identification model. The target artifact level identification model may determine a level of influence of artifacts in the candidate medical image on the image quality of the candidate medical image based on the target artifact attribute information.
Alternatively, the target artifact degree recognition model may recognize candidate medical images and divide regions of interest and regions of non-interest in the medical image to be processed. Wherein the region of interest may be a scanned region comprised in the medical image to be processed. For example, when the scanned region is a brain, the image to be processed includes image information corresponding to the brain and image information corresponding to a small portion of the neck. The target artifact level recognition model classifies a neck in the medical image to be processed as a region of non-interest and a brain in the medical image to be processed as a region of interest.
After determining the region of interest in the candidate medical image, the target artifact degree identification model may determine the degree of influence of the artifact in the candidate medical image on the image quality according to the position information of the region of interest in the candidate medical image, the attribute information of the region of interest, and the attribute information of the target artifact.
Illustratively, the scanned part corresponding to the candidate medical image is a brain, the target artifact degree identification model identifies the candidate medical image, brain tissues such as white matter, grey matter and the like in the candidate medical image are identified as regions of interest, and necks in the candidate medical image are identified as non-regions of interest. The target artifact degree model determines that a target artifact in the candidate medical image is a neck motion artifact according to the target artifact attribute information. Since the neck motion artifact has little effect on brain tissue, the target artifact level recognition model determines that the artifact in the candidate medical image has less effect on image quality.
Alternatively, the target artifact degree recognition model may determine the degree of influence of the artifact in the medical image to be processed on the image quality of the medical image to be processed according to the difference between the signal-to-noise ratio, the contrast, and the like of the medical image to be processed in which the target artifact exists and a set image quality threshold (for example, a signal-to-noise ratio threshold, a contrast threshold), and the like. If the signal-to-noise ratio, the contrast ratio and the like of the medical image to be processed with the target artifact are larger than the set image quality threshold value, the target artifact degree identification model determines that the artifact in the medical image to be processed has larger influence degree on the image quality of the medical image to be processed; if the signal-to-noise ratio, the contrast ratio and the like of the medical image to be processed, in which the target artifact exists, are smaller than the set image quality threshold value, the target artifact degree identification model determines that the artifact in the medical image to be processed has smaller influence degree on the image quality of the medical image to be processed.
Optionally, the target artifact degree identification model may determine a degree of influence of the artifact in the medical image to be processed on the image quality of the medical image to be processed according to a positional relationship between the target artifact and the scanned region. If the distance between the target artifact and the scanned part is smaller than a preset distance threshold, the target artifact degree identification model determines that the artifact in the medical image to be processed has a larger influence degree on the image quality of the medical image to be processed; if the distance between the target artifact and the scanned part is greater than or equal to a preset distance threshold, the target artifact degree identification model determines that the artifact in the medical image to be processed has smaller influence on the image quality of the medical image to be processed.
Optionally, the target artifact degree identification model may determine a degree of influence of the artifact in the medical image to be processed on the image quality of the medical image to be processed according to the area size of the target artifact. If the area of the target artifact exceeds a preset area threshold, the target artifact degree identification model determines that the artifact in the medical image to be processed has a larger influence on the image quality of the medical image to be processed; if the area of the target artifact is smaller than the preset area threshold value, the target artifact degree identification model determines that the artifact in the medical image to be processed has smaller influence on the image quality of the medical image to be processed.
Optionally, the target artifact degree identification model may determine a degree of influence of the artifact in the medical image to be processed on the image quality of the medical image to be processed according to the number of target artifacts. If the number of the target artifacts exceeds a preset number threshold, the target artifact degree identification model determines that the artifacts in the medical image to be processed have a larger influence on the image quality of the medical image to be processed; if the number of target artifacts is smaller than the preset number threshold, the target artifact degree identification model determines that the artifacts in the medical image to be processed have a smaller influence on the image quality of the medical image to be processed.
Alternatively, the target artifact level recognition model may determine the degree of influence of the artifact in the candidate medical image on the image quality according to the type of the target artifact. If the types of the target artifacts belong to the types of the unavoidable artifacts of the scanned part in the scanning process, the target artifact degree identification model determines that the influence degree of the artifacts in the medical image to be processed on the image quality of the medical image to be processed is smaller; if the types of the target artifacts belong to the types of the artifacts which can be avoided by the scanned part in the scanning process, the target artifact degree identification model determines that the artifacts in the medical image to be processed have a larger influence degree on the image quality of the candidate medical image.
Accordingly, the present invention also provides a computer readable storage medium storing at least one instruction executable by a processor, the at least one instruction implementing a medical image processing method as described above when executed by the processor.
In summary, the present invention provides a medical image processing method and a computer readable storage medium, including: acquiring a plurality of medical images to be detected; performing preliminary screening on a plurality of medical images to determine candidate medical images; and carrying out image quality evaluation on the candidate medical images to obtain quality evaluation results. According to the medical image processing method provided by the invention, before the quality evaluation of the medical image, the candidate medical image is determined through preliminary screening, so that the image quality can be rapidly screened, the calculation force is saved, and the detection efficiency of the image quality is improved.
The above description is only illustrative of the preferred embodiments of the present invention and is not intended to limit the scope of the present invention, and any alterations and modifications made by those skilled in the art based on the above disclosure shall fall within the scope of the appended claims.
Claims (10)
1. A medical image processing method, comprising:
acquiring a plurality of medical images to be detected;
performing preliminary screening on a plurality of medical images to determine candidate medical images;
and carrying out image quality evaluation on the candidate medical images to obtain quality evaluation results.
2. The medical image processing method of claim 1, wherein performing a preliminary screening of a plurality of the medical images to determine candidate medical images comprises:
segmenting a tissue region in at least one of the medical images;
setting a first threshold and a second threshold according to the segmented medical image respectively, and calculating an evaluation parameter, wherein the evaluation parameter is the ratio of the number of pixels with pixel values between the first threshold and the second threshold in the medical image to the number of pixels in the tissue region;
and performing preliminary screening on the medical images according to the relation between the evaluation parameters and a set threshold value to obtain candidate medical images.
3. The medical image processing method according to claim 2, wherein a maximum pixel value in a background of the medical image is extracted as a first threshold value and a minimum pixel value in the tissue region is extracted as a second threshold value.
4. The medical image processing method according to claim 2, further comprising identification of a body part in the medical image before segmenting the tissue region in the medical image.
5. The medical image processing method according to claim 4, wherein the identifying of the body part in the medical image is preceded by a preprocessing of the medical image, the preprocessing comprising a downsampling, filtering or normalizing process.
6. The medical image processing method according to claim 2, wherein when the evaluation parameter is greater than the set threshold value, the medical image is determined to be the candidate medical image.
7. The medical image processing method according to claim 1, wherein a machine learning model is used to evaluate the quality of the candidate medical image, the quality being characterized by at least one image attribute parameter of image integrity, image contrast, image signal to noise ratio, image resolution.
8. The medical image processing method according to claim 7, wherein the quality evaluation result includes normal, general, moderate abnormality, or severe abnormality.
9. The medical image processing method according to claim 1, wherein the medical image is obtained by scanning a scanning object by a scanner, the medical image processing method further comprising: and determining the state parameters of the scanner according to the quality evaluation result.
10. A computer-readable storage medium storing at least one instruction executable by a processor, the at least one instruction, when executed by the processor, implementing the medical image processing method according to any one of claims 1 to 9.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111652660.4A CN116434918A (en) | 2021-12-30 | 2021-12-30 | Medical image processing method and computer readable storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111652660.4A CN116434918A (en) | 2021-12-30 | 2021-12-30 | Medical image processing method and computer readable storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116434918A true CN116434918A (en) | 2023-07-14 |
Family
ID=87091197
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111652660.4A Pending CN116434918A (en) | 2021-12-30 | 2021-12-30 | Medical image processing method and computer readable storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116434918A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117808718A (en) * | 2024-02-29 | 2024-04-02 | 江西科技学院 | Method and system for improving medical image data quality based on Internet |
-
2021
- 2021-12-30 CN CN202111652660.4A patent/CN116434918A/en active Pending
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117808718A (en) * | 2024-02-29 | 2024-04-02 | 江西科技学院 | Method and system for improving medical image data quality based on Internet |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US7672517B2 (en) | Method for encoding image pixels a method for processing images and a method for processing images aimed at qualitative recognition of the object reproduced by one or more image pixels | |
| CN110506278B (en) | Target detection in hidden space | |
| US20190320934A1 (en) | Medical image acquisition with sequence prediction using deep learning | |
| US8588486B2 (en) | Apparatus and method for isolating a region in an image | |
| CN107545584B (en) | Method, device and system for positioning region of interest in medical image | |
| US7177453B2 (en) | Method and apparatus for partitioning a volume | |
| JP5562598B2 (en) | Image display apparatus, image display method, and magnetic resonance imaging apparatus | |
| US20240055126A1 (en) | Medical Image Segmentation Using Deep Learning Models Trained with Random Dropout and/or Standardized Inputs | |
| CN110807755A (en) | Plane selection using localizer images | |
| KR20190061041A (en) | Image processing | |
| US11969265B2 (en) | Neural network classification of osteolysis and synovitis near metal implants | |
| US11257211B2 (en) | Medical image processing apparatus, medical image processing system, and medical image processing method | |
| CN110288698B (en) | Meniscus three-dimensional reconstruction system based on MRI | |
| US8036730B1 (en) | Temporal magnetic resonance imaging | |
| US20210225491A1 (en) | Diagnostic image converting apparatus, diagnostic image converting module generating apparatus, diagnostic image recording apparatus, diagnostic image converting method, diagnostic image converting module generating method, diagnostic image recording method, and computer recordable recording medium | |
| CN109949288A (en) | Tumor type determines system, method and storage medium | |
| US20200022660A1 (en) | Systems and methods for cardiac triggering of an imaging system | |
| Lu et al. | Cardiac chamber segmentation using deep learning on magnetic resonance images from patients before and after atrial septal occlusion surgery | |
| CN116434918A (en) | Medical image processing method and computer readable storage medium | |
| CN115249279A (en) | Medical image processing method, medical image processing device, computer equipment and storage medium | |
| CN115700740A (en) | Medical image processing method, apparatus, computer device and storage medium | |
| CN115137342A (en) | Image acquisition medical system and method based on deep learning | |
| KR20200057463A (en) | Diagnostic Image Converting Apparatus, Diagnostic Image Converting Module Generating Apparatus, Diagnostic Image Recording Apparatus, Diagnostic Image Converting Method, Diagnostic Image Converting Module Generating Method, Diagnostic Image Recording Method, and Computer Recordable Recording Medium | |
| Chen et al. | “One-Shot” Reduction of Additive Artifacts in Medical Images | |
| KR20240037867A (en) | Learning method of medical image standardization network model and standardization method of medical images through MR image harmonization method based on cross-center style transfer using deep generative adversarial network |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |